## Multivariable: Derivatives of Polar Functions, Integrals of Polar Functions.

### August 4, 2010

This is going to be a short post about basic differentiating and integrating of multivariable functions. Remember, just because something is polar doesn’t mean that it’s not *fun! *But, if something is fun, then it rarely is polar. Just saying.

## Differentiating.

Obviously, if you want to take , then you differentiate normally. But say we have some function and we want to find ; note that this is slightly different from polar differentiating, and, in fact, will give us the “slope” of the polar plot if we look at it as if it were in rectangular coordinates.

Let’s derive this. Okay, so, what do we know about and in terms of polar coordinates? Well,

so, that’s nice. Now, can we manipulate to give us something nice? Let’s apply the chain rule:

Now, notice that is saying, “Take the derivative of the function with respect to .” What is this? Well, , and since may be a function of (indeed, it usually is written this way), we have

by applying the product rule, and similarly,

which allows us to simplify the expression above to:

This may not look like the simpliest expression, but it is much easier than converting a function from polar form back into rectangular form and differentiating.

Let’s take an example. Let’s take the function , below.

Now, look at this. It looks pretty cool, right? Yeah, let’s find of this, and see what’s what. First, let’s just calculate so we can replace it into the equation for the rectangular derivative.

and, now, I’ll be the first to admit that this is really an ugly formula. But, for now, let’s not reduce it, and let’s just plug in some things to see if we get something nice. First notice that at (this is the very bottom of the graph) we note that we should have a horizontal tangent line. In fact, plugging in our value for , we find that our derivative gives us 0! Nicely done, folks.

Let’s just do one more case, just to show we aren’t makin’ stuff up and everything should work out as expected. Suppose we have a circle with radius 2; that is, in polar form, we have a constant radius: .

In rectangular form, what is this? It’s just , right? Well, let’s see if we get the same derivatives doing this both ways. First,

So that our above formula gives us

which is certainly nice.

Now, let’s use implicit differentiation to solve this rectangular-wise.

which implies that

which, if we recall how to put this into polar form, is

which is exactly what we wanted! Good.

## Integrating.

To begin, let’s describe what we’ll be doing to integrate in polar coordinates. First, let’s think of what we might do to find an area of a circle. Let’s split it up into pizza pie slices, and then take the area of those: these are sectors, right? So, how big of a chunk is the sector of the circle? Let’s suppose that is in radians.

and so, the total area of the sector is

as we know from geometry, where is the degree of each slice. What is the degree of each slice? Well, at this point, since it’s a circle, it doesn’t really matter, but let’s try to make them really small. In fact, let’s make it so that they are infinitely small — in other words, let’s make them into ‘s. How do we sum up infinitely small things? Integrate!

will give us the area of the circle for some constant radius. Now, how about if we were to change the radius? Well, since we have , it’s almost like we’re integrating little bits of concentric circles.

On this graph, for example, this part of the sector (VERY exaggerated), we’re pretending that this part of the graph is a circle. As we move, the circle gets a little bigger or smaller. So, basically, if we have some equation we can integrate it by replacing for . We then get

for the area under a polar curve from to .

Now, be cautious here: if the graph “overlaps” itself, you’ll compute a much larger area than you wanted! Let’s give an example of this.

The function , as you recall, is a circle with radius 2, and we have

but, if we were to take

so what gives? It’s because in the latter one, we went around the circle *twice.* Yeah, so it’s kind of like looking at a spiral, and the first one goes around once and the second one goes around twice.

Now, let’s just look at slightly more difficult example. What about ?

Well, let’s integrate this.

and, as you know, the power-reducing formula for cosine squared gives us:

but, now, what are our limits of integration? Well, if we look at the graph, it’s going from to .

So, this reduces to

But how can we check this? Notice that this is, in fact, a circle, and that the farthest point is when cosine is equal to 1; in other words, this is a circle of diameter 3. This means the radius is equal to and, therefore, the area is . I’ll hold for gasps and applause.

## Observations and Notes.

This is essentially it. The hardest part is really finding out what the upper and lower limits of this type of equation is. For this, my best solution has been: use wolfram alpha. Seriously. I’m pretty lazy.

Since this type of thinking will generalize to, say, spherical and cylinderal coordinates, it would be easy to derive similar methods of differentiation and integration for these representations.

Maybe we’ll do these. Maybe.